Temporal Factors of EEG and Artificial Neural Network Classifiers of Mental Workload

B. N. Penaranda,Carryl L. Baldwin

DOI: https://doi.org/10.1177/1071181312561016

2012-09-01

Proceedings of the Human Factors and Ergonomics Society Annual Meeting

Abstract:The use of machine learning algorithms to classify Mental Workload (MW) from various neurophysiological measures is a growing trend in Human Factors research. Several teams have demonstrated that artificial neural networks (ANNs) can be employed to differentiate highly-dimensional neurophysiological data from participants performing tasks at different levels of cognitive demand in many experimental paradigms. Yet, in other cases, classifier performance was found not to exceed chance levels. One inescapable aspect of neurophysiological measures is the time course associated with data collection. This study directly examined the effect of time on classifier performance. Relative classifier performance values were compared for ANNs trained with EEG data from participants performing a verbal/spatial n-back task at varying load levels. For the vast majority of participants, classifiers trained with data from one session were ineffective when tested with data from a subsequent session. Similarly, classification performance suffered when training and testing tasks were incongruent. In contrast to the very high aggregate performance when trained and tested with data from the same session and same task (M = .81), cross-classification performance was consistently much worse (M = .49 – .53). The relative magnitude of the cross-session and cross-task “costs” to classification were compared, revealing a greater effect for session than task. The authors argue that the mere passage of time causes tonic changes in MW-related features of EEG that severely confound ANNs.